Correlation between Chiral Modifier Adsorption and Enantioselectivity in Hydrogenation Catalysis

Article abstract of DOI:10.1002/anie.201704880

Infrared absorption spectroscopy performed in situ at the solid–liquid interface revealed that the adsorption on platinum supported catalysts of 1-(1-naphthyl)-ethylamine, which is used as a chiral modifier in hydrogenation catalysis, occurs through the amine group, not the aromatic ring as is widely believed. Comparisons were performed against a set of related modifier compounds with targeted substitutions to help identify the key moiety involved in the adsorption. It was determined that neither naphthalene-based modifiers without amine groups nor those with tertiary amine moieties are capable of adsorbing on the metal surface to any significant extent. A direct correlation was also found between the ability of the amines to adsorb on the platinum surface and their performance as chiral modifiers that impart enantioselectivity to the hydrogenation of α-keto esters such as ethyl pyruvate.

Full text of DOI:10.1002/anie.201704880

A Journal of the Gesellschaft Deutscher Chemiker
Angewandte
International Edition Chemie
www.angewandte.org
Accepted Article
Title: Correlation Between Chiral Modifier Adsorption and
Enantioselectivity in Hydrogenation Catalysis
Authors: Yufei Ni, Alexander D. Gordon, Florisa Tanicala, and
Francisco Zaera
This manuscript has been accepted after peer review and appears as an
Accepted Article online prior to editing, proofing, and formal publication
of the final Version of Record (VoR). This work is currently citable by
using the Digital Object Identifier (DOI) given below. The VoR will be
published online in Early View as soon as possible and may be different
to this Accepted Article as a result of editing. Readers should obtain
the VoR from the journal website shown below when it is published
to ensure accuracy of information. The authors are responsible for the
content of this Accepted Article.
To be cited as: Angew. Chem. Int. Ed. 10.1002/anie.201704880
Angew. Chem. 10.1002/ange.201704880
Link to VoR: http://dx.doi.org/10.1002/anie.201704880
http://dx.doi.org/10.1002/ange.201704880

Angewandte Chemie International Edition
10.1002/anie.201704880
COMMUNICATION
Correlation Between Chiral
Modifier Adsorption and
Enantioselectivity in
Hydrogenation Catalysis
Yufei Ni, Alexander D. Gordon, Florisa Tanicala,
[
a]
and Francisco Zaera*
Abstract: By combining infrared absorption spectroscopy,
performed in situ at the solid-liquid interface, with the use of a set of
related compounds with targeted substitutions, the adsorption on the
surface of platinum supported catalysts of 1-(1-naphthyl)-ethylamine
(
NEA), used as a chiral modifier in hydrogenation catalysis, was
determined to occur via the amine group, not the aromatic ring as
widely believed. It was established that neither naphthalene-based
molecules without amine groups nor ones with tertiary amine
moieties are capable of adsorbing on the metal surface to any
significant extent. A direct correlation was then found between the
ability of the amines to adsorb on the Pt surface and their
performance as chiral modifiers, imparting enantioselectivity to the
hydrogenation of α-keto esters such as ethyl pyruvate.
Figure 1. Orito reaction and structure of the modifiers used in this study.
Chiral carbon atoms are indicated by asterisks.
The IR spectra of the modifiers listed in Figure 1, adsorbed
from 1 mM CCl
4
solutions onto a commercial 1 wt% Pt/SiO
2
catalyst, are displayed in Figure 2.
Only S-NEA, 1-
naphthylmethylamine (NMA), and quinoline (Q) adsorb on the
platinum surface to any significant extent; no measurable uptake
could be detected with either 1-ethylnaphthalene (EtN) or S-
N,N,-dimethyl-1-(1-naphthyl)-ethylamine (S-DNE). Unfortunately,
it is difficult to quantify the signals from these IR spectra to
calculate surface coverages for the adsorbed species, but it is
easy to see that, on average, the intensities of the most
prominent peaks in the IR traces for S-DNE and EtN are at least
one order of magnitude weaker than those for NEA, NMA and
quinoline (a variation not seen in the spectra of the pure
compounds). What is significant about this result is that,
although all of these compounds contain aromatic rings
In order to enhance the selectivity of heterogeneous catalysts, it
is imperative to develop an understanding between the nature of
the active sites and their reactivity. Recent advances in
spectroscopies that operate in-situ under catalytic conditions
have greatly facilitated the identification of structure-activity
correlations in processes involving gaseous reactants and
[
1-3]
products.
However, analogous research with solid-liquid
[
4, 5]
interfaces remains a challenge.
Here we discuss our results
from a study on the adsorption of chiral molecules from solutions
onto solid surfaces as it pertains to the bestowing of
enantioselectivity to hydrogenation catalysts. In the specific
system addressed, 1-(1-naphthyl)-ethylamine (NEA) is added
from solution to platinum catalysts to direct the chirality of the
hydrogenation of ethyl pyruvate (EtPy) to ethyl lactate (EtLac), a
prototypical example of the so-called Orito reaction (Figure 1,
2
(naphthalene, except for Q), only the ones with a NH moiety are
capable of bonding to the Pt surface. The common wisdom has
been that chiral modifiers adsorb on the metal surface via their
[
7-16]
aromatic ring,
but that does not appear to be the case here.
[
6, 7]
top).
By carrying out attenuated-total-reflection infrared
In principle, the species detected by IR could be bonded to the
acidic sites of the alumina support, but we have previously
shown that the same trend is seen for adsorption on a platinum
absorption spectroscopy (ATR-IR) experiments in-situ and
comparatively using a set of related compounds (Figure 1,
bottom), it was found that NEA bonds to the metal through the
nitrogen atom of its amine moiety, not via the aromatic ring as
[
17]
foil, where no oxide surfaces are present; only the behaviour
of quinoline is different between those two cases (it adsorbs on
[
7-16]
commonly believed.
those modifiers that contain the non-tertiary amine moiety
required for adsorption are capable of promoting
enantioselectivity in hydrogenation catalysis.
Moreover, it was determined that only
Pt/Al
2 3
O but not on the Pt foil). Blank experiments were also
carried out with the support alone to rule out this possibility.
[
a]
Mr. Yufei Ni, Dr. Alexander D. Gordon, Ms. Florisa Tanicala, and
Prof. Francisco Zaera
Department of Chemistry and UCR Center for Catalysis
University of California, Riverside, CA 92521, USA
E-mail: zaera@ucr.edu
This article is protected by copyright. All rights reserved.

Angewandte Chemie International Edition
10.1002/anie.201704880
COMMUNICATION
Figure 3. IR absorption spectra of NEA adsorbed on Pt surfaces under
different conditions. The trace for pure NEA is provided at the bottom for
reference.
The data show that NEA adsorbed from the gas phase
(
UHV) is mostly undisturbed, as the IR peaks are almost all seen
at the same frequencies as in the spectrum of the pure
[
18]
compound.
There are some differences in peak relative
intensities between the data for the low (0.5 L) and high (1.5 L)
coverages, reflecting a change in adsorption geometry, but all of
the peak frequencies are well matched between those two
traces. The traces for NEA adsorbed from the liquid solutions,
on the other hand, display significant changes in both peak
positions and peak relative intensities, a fact that highlights the
importance of characterizing the adsorption of these chiral
modifiers in situ at the solid-liquid interface to better emulate the
[
22]
reaction conditions.
-
1
Table 1. Assignment of IR peaks for adsorbed NEA. All values are in cm .
Figure 2. Attenuated total reflection infrared (ATR-IR) absorption spectra for
a
Relative intensities are given in parenthesis.
the modifiers in Figure 1, adsorbed from
4
1 mM CCl solutions onto a
commercial 1 wt% Pt/SiO catalyst. The spectra of the pure compounds are
provided as faded lines for reference (the scale for those is arbitrary, as the
peak intensities depend on the thickness of the films probed).
2
b
Mode
Liquid
Pt(111)
Pt Foil
Pt/Al
2
O
3
Pt/SiO
2
δ
δ
δ
oop(CH)ring
ip(CH)ring
1103 (s)
1101 (s)
1101 (m)
1119 (sh)
1163 (m)
1119 (vs)
1167 (vs)
1121 (s)
1169 (s)
Figure 2 also shows that the spectra of the adsorbed species
are quite different to those of the pure compounds. These
differences are explored in more detail for the case of S-NEA in
Figure 3, which displays IR absorption spectra for (from bottom
to top): (1) pure NEA (acquired in transmission mode); (2 and 3)
NEA adsorbed on a Pt(111) single crystal under ultrahigh
vacuum (UHV, acquired in single reflection-absorption, or
RAIRS, mode), after two exposures: 1.5 L (which leads to a
oop(CH)ring
1
203 (m)
δ
δ
ip(CH)ring
ip(CH)ring
1204 (m)
1204 (m)
1260 (m)
1201 (m)
1257 (m)
1200 (m)
1256 (m)
1217 (sh)
1240 (s)
1260 (s)
1252 (w)
1263 (sh)
1
327 (sh)
1327 (w)
1342 (m)
1325 (w)
1344 (m)
-
6
coverage close to monolayer saturation; 1 L = 1 x 10 Torr s)
δ(CH)amine
3 amine
δ (CH )
1339 (m)
1333 (m)
1379 (m)
1331 (m)
1377 (m)
[
18]
and 0.5 L (representative of the low-coverage limit); (4) NEA
[
17, 19]
1367 (s)
1369 (vs)
1396 (m)
1446 (s)
1381 (m)
1398 (m)
s
adsorbed from a 1 mM CCl
4
solution on a polished Pt foil
[
20]
(
also acquired in RAIRS mode but using a different setup);
1
401 (sh)
1400 (sh)
1408 (m)
and (5 and 6) NEA adsorbed from 1 mM CCl
4
solutions on 1
νip(CC) ring, s
1395 (s)
1412 (m)
wt% Pt/Al
2
O
3
and
1
wt% Pt/SiO
2
commercial catalysts,
[
21]
1448 (w)
1460 (w)
1447 (w)
1458 (w)
1450 (m)
1461 (sh)
respectively (acquired in ATR-IR mode). An assignment of
the peaks, based on previous studies, is provided in Table 1.
[
18]
δ
a
(CH
3
)
amine
1445 (vs)
1510 (vs)
1595 (vs)
1
514 (s)
ν
ν
ip(CC)ring, l
1511 (vs)
1597 (vs)
1526 (vs)
1524 (vs)
1539 (vs)
1
602 (s)
1601 (s),
1622 (s)
1595 (vs)
1610 (sh)
ip(CC)ring, s
1624 (s)
a
b
vs = very strong; s = strong; m = medium; w = weak; sh = shoulder.
δ = deformation; ν = stretching; ip = in plane; oop = out of plane, s =
symmetric; a = asymmetric.
First, it is clear that, in all the spectra for NEA adsorbed from
-
1
solution, the dominant peak is that around 1500 cm ,
corresponding to an in-plane deformation of the aromatic ring
along the long molecular axis. In addition, the corresponding
peaks due to the in-plane ring deformation along the short axis,
-
1
those around 1600 cm , display lower relative intensities than in
the neat liquid or when adsorbed under UHV. The blue shifts of
both peaks suggest an additional interaction with the solvent (or
a chemical transformation, see below), and the peak split seen
This article is protected by copyright. All rights reserved.

Angewandte Chemie International Edition
10.1002/anie.201704880
COMMUNICATION
in some cases indicates intermolecular interactions, as also
EtN, or NEA. In addition, none of the non-chiral modifiers
(quinoline, EtN, or NMA) add enantioselectivity to the catalyst,
as expected. However, no measurable enantioselective excess
[
18, 19]
observed with racemic mixtures.
Second, the peak from the asymmetric deformation of the
(
ee) was seen with S-DNE either, even though that modifier is
-
1
methyl moiety (1445-1460 cm ) of the ethylamine group is less
intense, in relative terms, than the corresponding symmetric
chiral. This is likely because S-DNE, a tertiary amine, does not
adsorb on the surface (Figure 2); a similar poor performance
-
1
(
umbrella) mode at ~1380 cm . In addition, the deformation
was reported with another (pyrrolidine-based) tertiary-amine
mode of the C–H bond next to the amine group is clearly visible
in the spectra of the species adsorbed from solution, at ~1330
[28]
NEA in the past.
On the other hand, both S-NEA and R-NEA
(
but not the racemic mixture) promote enantioselectivity during
-
1
cm . All together, it is this region of the spectra, which
corresponds mostly to vibrations within the ethyl moiety, that
shows the most differences in peak position between the
species adsorbed from solution and those dosed under UHV.
All of this is consistent with a model where NEA, when
adsorbing from solution, bonds to the Pt surface via the nitrogen
atom and has all of the substituents in the adjacent chiral carbon
the hydrogenation of EtPy, even if not as effectively as the
cinchona alkaloids (the reaction conditions were not optimized in
this study). There is a clear correlation between the ability of
these modifiers to adsorb on the surface of the catalyst and their
chiral modification behaviour: bonding through the amine group
may be central to their performance in catalysis.
(
the naphthalene and methyl groups and the extra hydrogen)
pointing away from the surface.
Table 2. Activity (% conversion) and enantioselective excess (% ee) for the
hydrogenation of ethyl pyruvate with a 1 wt% Pt/Al
2
O
3
catalyst as a function of
It may even be that adsorption requires the breaking of one
of the N–H bonds. Such dissociative adsorption, which would
likely be reversible, may explain why some of the deformation
modes of the C–H bonds in the aromatic ring (especially those
with an out-of-plane motion) are seen at much lower frequencies
than in the free molecule, and also account for the limited uptake
on the surface seen with DNE (a tertiary amine with no N–H
bonds). In fact, published IR data for the thermal decomposition
of N-methylaniline on Pt(111) (under UHV) show that the strong
a
the modifier used.
Modifier
% Conversion (±1.0)
% ee (±1.0)
0.2
No Modifier
Quinoline
EtN
15.8
12.3
13.1
23.2
25.8
12.3
12.5
10.5
54.5
68.8
24.5
25.9
0.3
0.9
-
1
NMA
0.7
peak for the ring deformation mode around 1530 cm shifts
toward higher frequencies, as seen here in going from vacuum
to the liquid-solid interface, upon dehydrogenation of the amine
S-DNE
0.4
[
23]
group.
detail.
We are at present exploring this possibility in more
S-NEA
24.4 (S)
29.0 (R)
0.2
R-NEA
The adsorption of NEA from solution onto Pt catalysts has
been previously studied with ATR-IR absorption spectroscopy by
Rac-NEA
Quinine
Quinidine
Cinchonine
Cinchonidine
a
[
16]
the Baiker group.
In their report, the broadening of the peaks
68.5 (R)
62.5 (S)
56.6 (S)
61.3 (R)
due to the vibrational modes of the naphthalene ring was taken
to indicate the interaction of that aromatic moiety with the
platinum surface. This is a conclusion different to that reached
by us. It should be said, however, that ATR-IR is not intrinsically
a surface-sensitive technique; the fact that the position of most
of the peaks seen in Baiker's data match those for NEA in
solution suggests that their ATR-IR spectra may correspond to
species in solution, not adsorbed on the surface. As mentioned
above, that possibility was ruled out in our work by carrying out
blank experiments without the catalyst and with the pure solid
support, without Pt. We have also attempted to study the
Reaction conditions: t = 20 min, T = 300 K, P(H
.85 µmol modifier, Pt:Modifier:EtPy molar ratio = 3:1:750, Solvent: 6 mL
acetic acid.
2
) = 1 bar, 50 mg catalyst,
0
adsorption of NEA from toluene (rather than CCl
) solutions, as Experimental Section
4
was done in the Baiker's case, but were unsuccessful: the high
solubility of NEA in toluene leads to very low steady-state
The IR absorption spectra were obtained using a Tensor 27 Bruker
Fourier-Transform infrared spectrometer (FTIR) and commercial
multiple-bounce attenuated total reflection (ATR) accessory purchased
from Pike Technologies and equipped with a 60-mm long Ge prism.
mg of the catalyst, either a 1 wt% Pt/Al (Alfa Aesar) or a 1 wt%
Pt/SiO (Sigma-Aldrich) solid, were dispersed on the surface of the prism,
and the ATR cell filled with the pure solvent (CCl , Sigma-Aldrich, 99.9%
purity). The catalysts were preconditioned by bubbling H gas (Liquid
Carbonic, 99.995% purity) for 60 min, after which a background IR scan
was taken. The cell was flushed with a 1 mM CCl solution of the
[
24]
a
coverages of adsorbed NEA on the Pt surface.
The activity and enantioselectivity of the chirally-modified
Pt/Al catalysts were also tested for the hydrogenation of EtPy
the Orito reaction in Figure 1). Table 2 summarizes the kinetic
3
O
3
2 3
O
2
(
2
4
data obtained with all of the modifiers in Figure 1 plus the four
most common cinchona alkaloids used in this reaction (quinine,
quinidine, cinchonine, and cinchonidine). An enhancement in
activity is seen with all the cinchona alkaloids, as reported in the
2
4
modifier (quinoline, >97% purity; EtN, >99% purity; S-DNE, 97% purity;
NMA, 97% purity, S-NEA, >99% purity; all from Sigma-Aldrich), and
[
25-27]
past,
and also with NMA and S-DNE, but not with quinoline,
This article is protected by copyright. All rights reserved.

Angewandte Chemie International Edition
10.1002/anie.201704880
COMMUNICATION
spectra were taken periodically until reaching surface saturation (~ 2 h).
The cell was then flushed with pure solvent, and another spectrum
acquired to differentiate between weakly- and strongly-bonded
adsorbates. The spectra were acquired by adding 256 scans taken with
[6]
B. Minder, M. Schuerch, T. Mallat, A. Baiker, T. Heinz, A. Pfaltz, J.
Catal. 1996, 160, 261.
[7]
[8]
A. Pfaltz, T. Heinz, Top. Catal. 1997, 4, 229.
G. Bond, P. A. Meheux, A. Ibbotson, P. B. Wells, Catal. Today 1991, 10,
371.
-
1
4
cm resolution, and were ratioed against the reference spectrum
recorded at the start of the experiment.
[9]
O. Schwalm, B. Minder, J. Weber, A. Baiker, Catal. Lett. 1994, 23, 271.
[
[
10] G. Vayner, K. N. Houk, Y.-K. Sun, J. Am. Chem. Soc. 2004, 126, 199.
11] M. Bartók, M. Sutyinszki, K. Balázsik, G. Szöllodblacsi, Catal. Lett.
For the catalytic measurements, the catalyst was first preconditioned via
2
005, 100, 161.
sequential treatments in Ar at room temperature for 90 min, O
for 120 min, and H at 350°C for 120 min, after which it was cooled down
under a H atmosphere and transferred to a two-necked flask and mixed
2
at 300°C
[
12] A. Kraynov, A. Suchopar, L. D'Souza, R. Richards, Phys. Chem. Chem.
Phys. 2006, 8, 1321.
2
2
[
[
13] H.-U. Blaser, M. Studer, Acc. Chem. Res. 2007, 40, 1348.
14] A. Taskinen, V. Nieminen, M. Hotokka, D. Y. Murzin, J. Phys. Chem. C
with the solvent (acetic acid, >99.7% purity, Sigma-Aldrich). Solutions of
the modifier and the reactant (EtPy, 98% purity, Sigma-Aldrich) were
2
007, 111, 5128.
2
added while bubbling H and stirring continuously using a magnetic bar.
[
15] V. Demers-Carpentier, G. Goubert, F. Masini, R. Lafleur-Lambert, Y.
Dong, S. Lavoie, G. Mahieu, J. Boukouvalas, H. Gao, A. M. H.
Rasmussen, L. Ferrighi, Y. Pan, B. Hammer, P. H. McBreen, Science
Small aliquots of the solution were taken periodically and analysed by
gas chromatography using a chiral column (DB-WAX, Agilent).
2
011, 334, 776.
[
[
16] F. Meemken, T. Steiger, M. C. Holland, R. Gilmour, K. Hungerbuhler, A.
Baiker, Catal. Sci. Technol. 2015, 5, 705.
Acknowledgements
17] A. D. Gordon, F. Zaera, Angew. Chem., Int. Ed. 2013, 52, 3453.
Financial assistance for this project was provided by a grant
from the US Department of Energy, Award # DE-SC0008703.
[18] I. Lee, Z. Ma, S. Kaneko, F. Zaera, J. Am. Chem. Soc. 2008, 130,
4597.
1
[
[
[
19] A. D. Gordon, S. Karakalos, F. Zaera, Surf. Sci. 2014, 629, 3.
20] J. Kubota, Z. Ma, F. Zaera, Langmuir 2003, 19, 3371.
21] F. Zaera, Chem. Soc. Rev. 2014, 43, 7624
Keywords: Enantioselectivity • Chiral Modifier • Structure-
Activity Relationships • IR Spectroscopy • Adsorption
[22] Z. Ma, I. Lee, F. Zaera, J. Am. Chem. Soc. 2007, 129, 16083.
[
[
[
23] K. Mudiyanselage, M. Trenary, Surf. Sci. 2009, 603, 3215.
24] Z. Ma, F. Zaera, J. Phys. Chem. B 2005, 109, 406.
[
[
[
1]
2]
3]
E. Gross, G. A. Somorjai, J. Catal. 2015, 328, 91.
25] J. L. Margitfalvi, P. Marti, A. Baiker, L. Botz, O. Sticher, Catal. Lett.
F. Tao, P. A. Crozier, Chem. Rev. 2016, 116, 3487.
1
990, 6, 281.
A. Chakrabarti, M. E. Ford, D. Gregory, R. Hu, C. J. Keturakis, S. Lwin,
Y. Tang, Z. Yang, M. Zhu, M. A. Bañares, I. E. Wachs, Catal. Today
[
[
[
26] C. LeBlond, J. Wang, J. Liu, A. T. Andrews, Y.-K. Sun, J. Am. Chem.
Soc. 1999, 121, 4920.
2
017, 283, 27.
27] N. J. Colston, R. P. K. Wells, P. B. Wells, G. J. Hutchings, Catal. Lett.
[
[
4]
5]
F. Zaera, Chem. Rev. 2012, 112, 2920.
2
005, 103, 117.
O. Björneholm, M. H. Hansen, A. Hodgson, L.-M. Liu, D. T. Limmer, A.
Michaelides, P. Pedevilla, J. Rossmeisl, H. Shen, G. Tocci, E. Tyrode,
M.-M. Walz, J. Werner, H. Bluhm, Chem. Rev. 2016, 116, 7698.
28] T. Heinz, G. Wang, A. Pfaltz, B. Minder, M. Schuerch, T. Mallat, A.
Baiker, J. Chem. Soc., Chem. Commun. 1995, 1421.
This article is protected by copyright. All rights reserved.

Angewandte Chemie International Edition
10.1002/anie.201704880
COMMUNICATION
COMMUNICATION
Chiral modification of Pt
hydrogenation catalysts by 1-(1-
naphthyl)-ethylamine (NEA) requires
adsorption via the amine nitrogen
atom, not bonding through the
aromatic ring as widely believed. The
Yufei Ni, Alexander D. Gordon, Florisa
Tanicala, and Francisco Zaera*
Page No. – Page No.
ability to adsorb requires NH
2
groups,
Correlation Between Chiral Modifier
Adsorption and Enantioselectivity in
Hydrogenation Catalysis
as determined by studying a family of
related compounds using in-situ
infrared absorption spectroscopy, and
correlates with the enantioselectivity
of the catalyst.
This article is protected by copyright. All rights reserved.